Microarray technology has become a powerful tool for generating and analyzing gene expression profiles. Microarray expression analysis, however, generally demands large amounts of RNA that are often not available (see Wang, et al., BioTechniques 34:394-400 (2003)). Several RNA amplification techniques have been developed to overcome this problem. These techniques, however, generally suffer from a phenomenon known as amplification bias (see, e.g., U.S. Pat. No. 6,582,906). In these cases, the amplified population of RNA molecules does not proportionally represent the population of RNA molecules existing in the original sample.
For example, in the method disclosed by Eberwine and colleagues (see, e.g., U.S. Pat. Nos. 5,545,522; 5,716,785; 5,891,636; 5,958,688; and 6,291,170), a compound oligonucleotide is utilized for the amplification, wherein the compound oligonucleotide is provided with both a T7 promoter and a primer. A cDNA copy is created of an initial mRNA transcript using the compound oliognucleotide, with subsequent second strand synthesis to create a cDNA that is double stranded. RNA amplification is conducted via the promoter portion of the compound oligonucleotide, with transcription proceeding off of the cDNA's second strand. Since the second strand is used for transcription, the Eberwine method produces amplified RNA that is antisense to the initial mRNA sequence.
The Eberwine method, however, introduces a 3′ bias during each of its steps due to the incomplete processivities (i.e., the inability of an enzyme to remain attached to a nucleic acid molecule) of the enzymes utilized and the positioning of the RNA polymerase promoter (see, e.g., U.S. Pat. No. 6,582,906 and U.S. Patent Publication US2003/0104432). For example, the compound oligonucleotide used to produce first strand cDNA places the promoter at the 5′ end of the cDNA, which corresponds to the 3′ end of the message. This coupled with the inability of RNA polymerase to complete transcription of some templates (due perhaps to long polyA tail regions or interference from secondary and tertiary structures in the template) can result in a 3′ bias in the amplified antisense RNA population. In addition, if second strand cDNA synthesis by DNA polymerase is incomplete, these cDNAs will lack functional promoters, resulting in a reduced representation of the original RNA molecule (or possibly a complete absence) in the amplified population.